Stefan Esser writes:

> I'm using the same drive in my system, and never
> observed that behaviour. The cause of the command
> failure seems to be a QUEUE FULL condition, as you
> write below.

> Since only 4 tags are generally used, this is an
> "impossible" situation. The drive supports more
> than ten times as many simultanous commands ...

> Could you try the latest driver version as found
> in a FreeBSD-current source tree ?

It doesn't compile out-of-the-box:

cc  -c -O6 -Werror -I. -I../../../../arch -I../../../.. -I../../../../sys -DI586_
CPU -DXSERVER -DUCONSOLE -DINSECURE -DMACHINE_NONCONTIG -DTIMEZONE="0" -DDST="0" 
-DSWAPPAGER -DVNODEPAGER -DDEVPAGER -DDDB -DDIAGNOSTIC -DKTRACE -DSYSVMSG -DSYSVS
EM -DSYSVSHM -DCOMPAT_NOMID -DCOMPAT_10 -DCOMPAT_43 -DTCP_COMPAT_42 -DCOMPAT_SVR4
 -DCOMPAT_LINUX -DCOMPAT_FREEBSD -DUSER_LDT -DLKM -DFFS -DMFS -DNFSCLIENT -DNFSSE
RVER -DCD9660 -DMSDOSFS -DFDESC -DFIFO -DUNION -DGATEWAY -DINET -DGENERIC -D_KERN
EL -Di386  ../../../../dev/pci/ncr.c
In file included from ../../../../dev/pci/ncr.c:185:
../../../../sys/sysctl.h:186: field `e_vm' has incomplete type
../../../../dev/pci/ncr.c:1269: parse error before `&'
cc1: warnings being treated as errors
../../../../dev/pci/ncr.c:1269: warning: data definition has no type or storage c
lass
../../../../dev/pci/ncr.c: In function `ncr_attach':
../../../../dev/pci/ncr.c:3249: warning: passing arg 3 of `pci_map_int' from inco
mpatible pointer type
*** Error code 1

I've looked at the diffs between the NetBSD-1.1 and FreeBSD-current versions,
and they are sufficiently plentiful that I strongly suspect it would be a bigger
job to port it than I have time to undertake right now.  Also, there have been
some changes between NetBSD-1.1 and NetBSD-current, so when I get the chance,
I will probably try upgrading to NetBSD-current (whose ncr.c does not seem to
incorporate your recent changes).

> You may want to disable tags as a workaround.

OK, I'm now running a kernel with tags disabled.  It's been running without
a hitch for about 12 hours, but I won't be able to have a high degree of
confidence that that has fixed the problem until it has been running for
a week or so.

> Please let me know, whether the system works 
> reliably with tags disabled, and if you can

Will do.

> reliably with tags disabled, and if you can
> build a kernel with the most recent driver
> code, whether this improves things ...

See above.

Thanks very much for the suggestions and for the quick response.

--Jim Bernard